c2f · PyPI

No project description provided

Project description

Is Python an interpreted or compiled language? Trick question. It’s actually both. With tools like Cython, we can take the compilation step further and remove the interpreter loop almost entirely. Cython produces binaries much like C++, Go, and Rust do. Now with GitHub Actions the cross-platform build and release process can be automated for free for Open Source projects. This is an enormous opportunity to make the Python ecosystem 20-50% faster with a single pull request. This lightning talk walks through a GitHub workflow that publishes Cython-optimized wheels to PyPI. Discover how Cython can turbo-charge your Python code and GitHub Actions can simplify your cross-platform release process for free.

SF Python Holiday Party 2019

Grant Jenks
Dec. 4, 2019
grantjenks.com/docs/cython-for-all
github.com/grantjenks/python-c2f
pypi.org/project/c2f/#files

Is Python interpreted or compiled?

.
.
.
?
.
.
.
?
.
.
.
?

c2f.py

"Celsius to Fahrenheit Library"

def convert(celsius: float) -> float:
    "Convert Celsius to Fahrenheit"
    fahrenheit = celsius * 1.8 + 32
    return fahrenheit

c2f.cpython-38.pyc

>>> import c2f
>>> dis.dis(c2f.convert)
  6           0 LOAD_FAST                0 (celsius)
              2 LOAD_CONST               1 (1.8)
              4 BINARY_MULTIPLY
              6 LOAD_CONST               2 (32)
              8 BINARY_ADD
             10 STORE_FAST               1 (fahrenheit)
  7          12 LOAD_FAST                1 (fahrenheit)
             14 RETURN_VALUE

setup.py

from setuptools import setup
from Cython.Build import cythonize

setup(
    name='c2f',
    version='0.0.0',
    py_modules=['c2f'],
    ext_modules=cythonize('c2f.py'),
)

c2f.c

$ cython c2f.py

static PyObject * __pyx_convert(double __pyx_v_celsius)
{
  double __pyx_v_fahrenheit;
  PyObject *__pyx_r = NULL;
  __pyx_v_fahrenheit = ((__pyx_v_celsius * 1.8) + 32.0);
  __pyx_r = PyFloat_FromDouble(__pyx_v_fahrenheit);
  return __pyx_r;
}

c2f.so

$ python setup.py bdist_wheel

___pyx_convert:
push   rbp
mov    rbp, rsp
sub    rsp, 16
movsd  xmm0, qword ptr [rbp - 8]
mulsd  xmm0, qword ptr [rip + 1379]
addsd  xmm0, qword ptr [rip + 1379]
call   502 <PyFloat_FromDouble ...>
add    rsp, 16
pop    rbp
ret

.github/workflows/release.yml

name: release
on:
  push:
    tags:
      - v*
jobs:
  build-linux-cp38:
    runs-on: ubuntu-latest
    container: quay.io/pypa/manylinux2014_x86_64
    steps:
    ...

Matrix Build

build-macos:
  runs-on: macos-latest
  strategy:
    max-parallel: 4
    matrix:
      python-version: [3.5, 3.6, 3.7, 3.8]
  steps:
  ...

Mac Build Steps

- name: Set up Python ${{ matrix.python-version }} x64
  uses: actions/setup-python@v1
  with:
    python-version: ${{ matrix.python-version }}
    architecture: x64

- name: Install package dependencies
  run: pip install cython wheel

- name: Build binary wheel
  run: python setup.py bdist_wheel

Linux auditwheel Tool

- name: Build binary wheel
  run: /opt/python/cp38-cp38/bin/python setup.py bdist_wheel

- name: Apply auditwheel for manylinux wheel
  run: auditwheel repair -w dist dist/*

- name: Remove linux wheel
  run: rm dist/*-linux_x86_64.whl

Windows Build Steps

- name: Download Build Tools for Visual Studio 2019
  run: Invoke-WebRequest -Uri https://aka.ms/vs/16/rel...

- name: Run vs_buildtools.exe install
  run: ./vs_buildtools.exe --quiet --wait --norestart ...

Store Build Artifacts

- name: Archive dist artifacts
  uses: actions/upload-artifact@v1
  with:
    name: dist-macos-${{ matrix.python-version }}
    path: dist

Source Distribution

upload:
  needs: [build-linux-cp35, ...]
  runs-on: ubuntu-latest
  steps:
  ...
  - name: Install dependencies
    run: pip install -r requirements.txt

  - name: Create source dist
    run: python setup.py sdist

Stage Binary Wheels

- name: Stage linux 3.8
  uses: actions/download-artifact@v1
  with:
    name: dist-linux-3.8
- run: mv -v dist-linux-3.8/* dist/

- name: Stage macos 3.8
  uses: actions/download-artifact@v1
  with:
    name: dist-macos-3.8
- run: mv -v dist-macos-3.8/* dist/
...

Upload with Twine

- name: Upload with twine
  env:
    TWINE_USERNAME: ${{ secrets.TWINE_USERNAME }}
    TWINE_PASSWORD: ${{ secrets.TWINE_PASSWORD }}
  run: |
    ls -l dist/*
    pip install twine
    twine upload dist/*

Cythonize all the Things!

PLEASE STEAL THE CODE!

grantjenks.com/docs/cython-for-all
github.com/grantjenks/python-c2f
pypi.org/project/c2f/#files

Cythonize all the Things!

Appendix

Dumping Assembly

$ gcc -g -shared -pthread -fPIC -fwrapv -O2 -Wall -fno-strict-aliasing -I/Library/Frameworks/Python.framework/Versions/3.8/include/python3.8 -L/Library/Frameworks/Python.framework/Versions/3.8/lib -o c2f.so c2f.c -lpython3.8
$ objdump -S -df=___pyx_pw_3c2f_1convert c2f.so

Git Tagging

$ git tag -a v0.0.2 -m v0.0.2
$ git push
$ git push --tags

Project details

Release history Release notifications | RSS feed

This version

1.0.5

Dec 6, 2019

1.0.4

Dec 5, 2019

1.0.3

Dec 5, 2019

1.0.0

Dec 5, 2019

0.0.2

Dec 4, 2019

0.0.1

Dec 4, 2019

0.0.0

Dec 4, 2019

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

c2f-1.0.5.tar.gz (28.5 kB view details)

Uploaded Dec 6, 2019 Source

Built Distributions

c2f-1.0.5-cp38-cp38-win_amd64.whl (16.7 kB view details)

Uploaded Dec 6, 2019 CPython 3.8 Windows x86-64

c2f-1.0.5-cp38-cp38-manylinux2014_x86_64.whl (49.3 kB view details)

Uploaded Dec 6, 2019 CPython 3.8

c2f-1.0.5-cp38-cp38-manylinux1_x86_64.whl (49.3 kB view details)

Uploaded Dec 6, 2019 CPython 3.8

c2f-1.0.5-cp38-cp38-macosx_10_13_x86_64.whl (13.8 kB view details)

Uploaded Dec 6, 2019 CPython 3.8 macOS 10.13+ x86-64

c2f-1.0.5-cp37-cp37m-win_amd64.whl (16.6 kB view details)

Uploaded Dec 6, 2019 CPython 3.7m Windows x86-64

c2f-1.0.5-cp37-cp37m-manylinux2014_x86_64.whl (47.1 kB view details)

Uploaded Dec 6, 2019 CPython 3.7m

c2f-1.0.5-cp37-cp37m-manylinux1_x86_64.whl (47.1 kB view details)

Uploaded Dec 6, 2019 CPython 3.7m

c2f-1.0.5-cp37-cp37m-macosx_10_13_x86_64.whl (13.7 kB view details)

Uploaded Dec 6, 2019 CPython 3.7m macOS 10.13+ x86-64

c2f-1.0.5-cp36-cp36m-win_amd64.whl (16.5 kB view details)

Uploaded Dec 6, 2019 CPython 3.6m Windows x86-64

c2f-1.0.5-cp36-cp36m-manylinux2014_x86_64.whl (46.1 kB view details)

Uploaded Dec 6, 2019 CPython 3.6m

c2f-1.0.5-cp36-cp36m-manylinux1_x86_64.whl (46.1 kB view details)

Uploaded Dec 6, 2019 CPython 3.6m

c2f-1.0.5-cp36-cp36m-macosx_10_13_x86_64.whl (13.6 kB view details)

Uploaded Dec 6, 2019 CPython 3.6m macOS 10.13+ x86-64

c2f-1.0.5-cp35-cp35m-manylinux2014_x86_64.whl (45.4 kB view details)

Uploaded Dec 6, 2019 CPython 3.5m

c2f-1.0.5-cp35-cp35m-manylinux1_x86_64.whl (45.4 kB view details)

Uploaded Dec 6, 2019 CPython 3.5m

c2f-1.0.5-cp35-cp35m-macosx_10_13_x86_64.whl (13.5 kB view details)

Uploaded Dec 6, 2019 CPython 3.5m macOS 10.13+ x86-64

File details

Details for the file c2f-1.0.5.tar.gz.

File metadata

Download URL: c2f-1.0.5.tar.gz
Upload date: Dec 6, 2019
Size: 28.5 kB
Tags: Source
Uploaded using Trusted Publishing? No
Uploaded via: twine/3.1.1 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.40.1 CPython/3.8.0

File hashes

Hashes for c2f-1.0.5.tar.gz
Algorithm	Hash digest
SHA256	`32a37e81f9011eda171b1e18efcff6f71a87642ce73f9cdae1e8ad885df2d24b`
MD5	`419e2aebfb117470ec3c104dedc544a9`
BLAKE2b-256	`38a40e62e09b7caeb357e7f0a752a560e64f59d0b0d00f1845020a2a3fb9a3a7`

Algorithm	Hash digest
SHA256	`8b7d71c8995a1192f494a8e45a0a523eddd869290b1933ef53d681ba68331cc3`
MD5	`bb6a1844a992e0cfb1a5e638c65c4df5`
BLAKE2b-256	`8c1997cd9261f66a6fc4a8ea65c4ca4506c4f7097fec2422c7f8e82101f8ee50`

Algorithm	Hash digest
SHA256	`b35ae1fb5879d1afbe5b2e7bfb714edd8fc083dc5a4b02e12212f791f6fb5775`
MD5	`10b50506d80580ba4ecb6eeded56851f`
BLAKE2b-256	`78c9501ed8f49eb7ab8a6d83fdd79bd963f98ce68e2aae67958b8af9efab321d`

Algorithm	Hash digest
SHA256	`af22ee0b125a11e952a5b0c4beb38608c8a6bee47c3d8ec9fce278b317c4e543`
MD5	`aaf8e214a8369c5839f713b7f22d4479`
BLAKE2b-256	`d5c6dee61af712f188f7893773e1ac37b2c256df9710dfc1c03174ed8da11913`

Algorithm	Hash digest
SHA256	`592d56360ec5fa7525c8874b2a8ba3e4d91e347152be296209378e4c7e5fe75e`
MD5	`f589dc56dd8dfe18ac700e398cf5c0ec`
BLAKE2b-256	`fdceb1338deadbaadac1dab0edf5b0e5c88cf497bf8d691eba7765cbf93713ad`

Algorithm	Hash digest
SHA256	`5f007efd3d1e01fb77c5413f19532d0d2b90051d51b73f4f87bb4927c73bafaa`
MD5	`df976dafdcd352496aa5116a544da3f2`
BLAKE2b-256	`92bdfc05400a4b90ddb555bd0c0cf29795d533e8be0e9ae3edad2929beee85b2`

Algorithm	Hash digest
SHA256	`62fc80307d7058f1ab197cfcc0414bcaae64674e9ef86909c24900e4646e4948`
MD5	`a6ccd36f2e65ae0191d333ddab185b6f`
BLAKE2b-256	`d31f042cca0538e4ec7d174bbf3b3f7a3cc7681d1c8ba03b3cd9b11426c9f98b`

Algorithm	Hash digest
SHA256	`e55252cc0da11995c3ec1ed9893d49f89d21a420c5c312503d9a0b03002e4b99`
MD5	`e90b80376ac82a1d7f032b29437a78fc`
BLAKE2b-256	`3c98b241e9c4e10dea2e7c552a25eebc8b83fb2b20a4607d7c3aa7869bf1f119`

Algorithm	Hash digest
SHA256	`cefd1ece181fda3dab69758f4dc479fe00c9cbb66ed0c31531b03a77d4bc5615`
MD5	`d3f0c3e42d012ee642b888f7d8bd51eb`
BLAKE2b-256	`b6754edf497ecf0023e75f477c4875311a0ac321cc4b5ecb3f80988644e03f8d`

Algorithm	Hash digest
SHA256	`921d4ec85ab00660d062169ad66907eb32f4e1ec97f3b123b87727effeb37194`
MD5	`6b77512af8d23097edd9f9dd02577dde`
BLAKE2b-256	`ec73e67420ed0dc962a2cf7f6db039ed7f1869a39b6474d0ba263bbd8e7c7350`

Algorithm	Hash digest
SHA256	`2d0828f6ec2325b262df735d24fe759d4468506f4afad2695cd9ca82065cab90`
MD5	`7431def9474b812973f84f3bd1ddc761`
BLAKE2b-256	`703f4a949a5ffb058039fc5c32a175ad26c8e088f5dbd04978e961eb7f5145ae`

Algorithm	Hash digest
SHA256	`e8107bd90fd33d505bc116f0499431969e55b7fc489bb45d2514a4931605d2c8`
MD5	`55bed57307c989d70fe0744cbdbdfdfb`
BLAKE2b-256	`b31074d162ca7b7f65e714e957a35ee1f9b0fd293fbcd87989e31e3a23103679`

Algorithm	Hash digest
SHA256	`7177a2034f173efda0cad6a34353f61cb70e2a81124f60942405c0e811210d62`
MD5	`0e8c04aae65cdafea32d0a54c03845ac`
BLAKE2b-256	`6e3817f6846d92af4e00f2a38f0e03fabeb1305c6c6dd579671e0df9f4f6dd18`

Algorithm	Hash digest
SHA256	`55d1687e7f585f09273ad3204f6d48c36b590b1232226b22575820e4e6fbf0f8`
MD5	`4b001883e8b50fe85246944704b02877`
BLAKE2b-256	`e3586efd27bd8d6df65702c91637478d41f3fbec703da974baa6aa1c0a8d15ac`

Algorithm	Hash digest
SHA256	`c40c5947b8a9ae50e16191cc8281ffadfec52d901289d79fd5ec42f025b67f17`
MD5	`527cd3180b3906ed1ab2b773e884db29`
BLAKE2b-256	`0d608e3f24bb3d34ace4843a27d4dde16212e6bf89e2e9b0cdad7388823a5fcd`

Algorithm	Hash digest
SHA256	`219cd3324f836703df7bb3ca89d4ea2e5b29584c03474905b34804b930de14fe`
MD5	`765a6967c38b6c4e2e62e0cdaa64280a`
BLAKE2b-256	`5db17ab342c928dc157b521bd73459aeb3ee8e698e28a30afaf6a4ff39c372bb`

c2f 1.0.5

Navigation

Verified details

Maintainers

Unverified details

Project description

SF Python Holiday Party 2019

Is Python interpreted or compiled?

c2f.py

c2f.cpython-38.pyc

setup.py

c2f.c

c2f.so

.github/workflows/release.yml

Matrix Build

Mac Build Steps

Linux auditwheel Tool

Windows Build Steps

Store Build Artifacts

Source Distribution

Stage Binary Wheels

Upload with Twine

Cythonize all the Things!

Appendix

Dumping Assembly

Git Tagging

Project details

Verified details

Maintainers

Unverified details

Release history Release notifications | RSS feed

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata